%auto-ignore

Changes made to paper:

* Inclusion of figure showing the simple volatile site complexity (as
used by C. Adami) as a function of speciation time. Altered results
section to describe this.

* Changed title to reflect the interim nature of this report

* Condensed section decsribing the two methods of searching genotype
space, as results reported don't use these methods (future ones
will though). This brings the paper within the 5 page limit.

Comments to referees

Ref 71. Some empirical results added. I'm not sure what the referee
means by saying this approach is highly biased. I do assume that the
bulk of neutral variants are connected within a single site mutation
neutral network, something that does appear to be the case in Tierra
(although by no means 100% certain). If this were not the case, then I
could see that this technique would be very biased,

Ref 66. Changed title to hopefully something a little less
"expectation raising". As to whether mathematical computations, such
as those performed by Avida, are important or not, my view is that
they are not, especially wrt to Tierra. We shall have to agree to
disagree on that one. I would also dispute that my method does not
discriminate between pairs which differ in ecological interactions,
as long as the overall fitness and cpu costs are the same. It very
clearly does, as the technique considers these quantities pair-wise
between all the players in the ecosystem.

Ref 119. I beg to differ with this referee who claims  coevolution
will grind to a halt in the absence of information. This seems far
from obvious to me, and is precisely why I'm interested in studying a
coevolving system that has some degree of tractibility. It may be that
a source of information is required in order to see complexity
increase (this seems plausible). In which case, Tierran evolution is
perhaps one long plateau of complexity, but churning through novel
organisms. I await my experimental results.
